Geoscience Reference
In-Depth Information
variables, as was traditionally considered. Significant differences exist in the
conceptual and statistical methods employed for the estimation of the DMR.
There are at least five different patterns analysed under the umbrella of
DMR, and not all mechanisms are equally appropriate for each of them
(see Maxwell and Jennings, 2006; Reuman et al., 2008; White et al., 2007 ). In
addition, recent advances in the estimation of parameters indicate that the
more widely used approaches are probably biased in parameter estimation,
because of a poor use of available information, violation of statistical
assumptions, and poor consideration of alternative models ( Jennings et al.,
2007; Packard et al., 2010; Reuman et al., 2009; White et al., 2008 ). As a
consequence, particular attention needs to be devoted to methodological
issues related to the inferred patterns.
In this chapter, we attempt to contribute to the advancement in the connec-
tion between food web structure and DMR, and also identify several methodo-
logical advances recently developed and evaluated. For this purpose, we briefly
introduce the different patterns considered within DMR and highlight their
connection to gape limitation. We then review the current state of knowledge
regarding the methodological issues related to the analysis of DMR for both
univariate (frequency distribution of body sizes) and bivariate relationships
(densities and body sizes). Further, we propose the use of some statistics to
detect changes in the DMR across the range of body sizes studied. Finally, with
both theoretical and methodological considerations in mind, we evaluate the
alternative DMRs in a metacommunity composed of 18 local communities
from which multiple species were measured at the individual level.
II. DENSITY-MASS FROM DIFFERENT ANGLES
It has long been suggested that a major source of variation between patterns
reported for different ecosystems could be the nature of the data used (e.g.
Cotgreave, 1993 ). In fact, White et al. (2007) highlighted that there are four
interrelated DMRs, a distinction rarely recognized in studies that typically
focus in just one of them ( Table 1 ). These patterns are the cross-community
scaling (CCSR) or self-thinning, global size-density relationship (GSDR),
individual size distribution (ISD) or size spectrum, and local size-density
relationship (LSDR; see White et al., 2007 ). In addition, the species mean-
size distribution (SMSD) has been proposed as a fifth scaling pattern,
connected to the ISD and the LSDR (see Reuman et al., 2008 ).
Self-thinning or, more generally, the CCSR is expected to reflect a size-
dependent competition generated by a limiting factor such as space or
resources, which grow in demand as the size of individuals increases
( Westoby, 1984; White et al., 2007; Yoda et al., 1963 ). It is expressed as a
negative relationship between the mean size of
individuals and total
Search WWH ::




Custom Search